文章基本信息

标题：Enhancement of ICA Technique for Change Detection in Hyperspectral Images
本地全文：下载
作者：Ruby Kamboj ; Manju Ahuja
期刊名称：International Journal of Research in Management, Science & Technology
印刷版ISSN：2321-3264
出版年度：2016
卷号：4
期号：2
出版社：Prannath Parnami Institute of Management & Technology, Hisar
摘要：The recent availability of remotely sensed high dimensional Hyperspectral image (HSI) has fostered the development of change detection techniques which are able to interpret such high-resolution data in various different applications contexts. Most of the researchers have been focussed on the Principal Component Analysis (PCA) technique of change Detection whereas it suffers from several drawbacks in case of high resolution HSI image. The reason behind these lacks is its principle of second order statistics. It fails to extract minute changes that could have occurred at the scene. Independent Component Analysis (ICA) is a technique based on higher order statistics and is capable of extracting minor changes that could have occurred from higher resolution HSIs. The aim of this paper is to enhance the performance of the FastICA algorithm for Hyperspectral image especially for performing change detection. The objective is achieved by tuning several parameters of the algorithm to such a combination that results in higher efficient and less complex algorithm for detecting changes. The paper is divided into six sections namely introduction, problem analysis, proposed method, implementation, result and discussions and performance analysis
关键词：Change Detection; Hyperspectral image; Principal Component Analysis (PCA); ICA; FastICA; Spectral signatures.